ORIGINAL RESEARCH ARTICLE published: 18 December 2013 doi: 10.3389/fncel.2013.00257 Tumor necrosis factor alpha maintains denervation-induced homeostatic of mouse dentate granule cells

Denise Becker, Nadine Zahn,Thomas Deller and Andreas Vlachos* Institute of Clinical Neuroanatomy, Center, Goethe-University Frankfurt, Frankfurt, Germany

Edited by: Neurons which lose part of their input respond with a compensatory increase in excitatory Arianna Maffei, State University of synaptic strength. This observation is of particular interest in the context of neurological New York Stony Brook, USA diseases, which are accompanied by the loss of neurons and subsequent denervation Reviewed by: of connected brain regions. However, while the cellular and molecular mechanisms of Carlos D. Aizenman, Brown University, USA pharmacologically induced homeostatic synaptic plasticity have been identified to a certain David Stellwagen, McGill University, degree, denervation-induced homeostatic synaptic plasticity remains not well understood. Canada Here, we employed the entorhinal denervation in vitro model to study the role of tumor Alanna Watt, McGill University, α Canada necrosis factor alpha (TNF ) on changes in excitatory synaptic strength of mouse dentate granule cells following partial deafferentation. Our experiments disclose that TNFα is *Correspondence: Andreas Vlachos, Institute of Clinical required for the maintenance of a compensatory increase in excitatory synaptic strength Neuroanatomy, Neuroscience Center, at 3–4 days post lesion (dpl), but not for the induction of synaptic scaling at 1–2 dpl. Goethe-University Frankfurt, Furthermore, laser capture microdissection combined with quantitative PCR demonstrates Theodor-Stern Kai 7, Frankfurt 60590, α Germany an increase in TNF -mRNA levels in the denervated zone, which is consistent with our e-mail: [email protected] previous finding on a local, i.e., layer-specific increase in excitatory synaptic strength at 3–4 dpl. Immunostainings for the glial fibrillary acidic protein and TNFα suggest that astrocytes are a source of TNFα in our experimental setting. We conclude that TNFα- signaling is a major regulatory system that aims at maintaining the homeostatic synaptic response of denervated neurons.

Keywords: entorhinal cortex lesion, homeostatic synaptic scaling, astrocytes, brain injury, organotypic slice cultures

INTRODUCTION TNFα affects synaptic strength (Beattie et al., 2002; Stellwagen Homeostatic synaptic plasticity is a slow adaptive mechanism et al., 2005; Leonoudakis et al., 2008; Santello et al., 2011; He et al., which allows neurons to adjust their synaptic strength to per- 2012; Pribiag and Stellwagen, 2013), its precise role in synaptic turbations in network activity. It aims at keeping the firing plasticity remains controversial. Recently experimental evidence rate of neurons within a dynamic range and is considered has been provided that TNFα may act as a permissive rather fundamental for the normal functioning of the central ner- than instructive factor (Steinmetz and Turrigiano,2010). Likewise, vous system. Hence, in response to a prolonged reduction its impact on synaptic plasticity under pathological conditions in network activity neurons increase (“scale up”) the strength remains not well understood (for a recent review on the role of of their excitatory (Turrigiano, 2012; Vitureira et al., TNFα in synaptic plasticity see Santello and Volterra, 2012). 2012; Wang et al., 2012; Davis, 2013). Using entorhinal den- Here, we studied the role of TNFα in denervation-induced ervation in vitro we recently showed that homeostatic synaptic synaptic plasticity using mature (≥18 days in vitro; div) entorhino- strengthening of excitatory synapses is observed in denervated hippocampal slice cultures (Del Turco and Deller, 2007). In these neuronal networks (Vlachos et al., 2012a, 2013a,b). This obser- cultures the axonal projection from the entorhinal cortex (EC) to vation indicates that homeostatic synaptic responses could play the outer molecular layer (OML) can be transected by removing an important role in a broad range of neurological diseases, the EC from the culturing dish (e.g., Vlachos et al., 2013a). This which are accompanied by the loss of central neurons and sub- leads to the partial deafferentation of dentate granule cells in the sequent denervation of connected brain regions. However, the OML, without directly damaging the dentate gyrus (DG; Müller molecular pathways involved in the regulation of denervation- et al., 2010). Using pharmacological and genetic approaches we induced homeostatic synaptic plasticity remain incompletely provide experimental evidence that denervation-induced synap- understood. tic plasticity is divided into a TNFα-independent early phase One of the factors suggested to control homeostatic synaptic [1–2 days postlesion (dpl)] and a TNFα-dependent late phase scaling following prolonged blockade of sodium channels with (3–4 dpl). Astrocytes seem to be a major source of TNFα in our tetrodotoxin (TTX) or pharmacological inhibition of ionotropic experimental setting. These results suggest an important role for glutamate receptors, is the pro-inflammatory cytokine TNFα TNFα in maintaining synaptic scaling responses in denervated (Stellwagen and Malenka, 2006). While it has been shown that neuronal networks, which could be of relevance in the context of

Frontiers in Cellular Neuroscience www.frontiersin.org December 2013 | Volume 7 | Article 257 | 1

“fncel-07-00257” — 2013/12/16 — 15:26 — page1—#1 Becker et al. TNFα and denervation-induced synaptic plasticity

neurological diseases in which neuronal death and denervation Germany) while recording evoked excitatory postsynaptic currents occur. from individual granule cells (Figure 2A; c.f. Vlachos et al., 2012a).

MATERIALS AND METHODS DRUG TREATMENTS PREPARATION OF SLICE CULTURES Denervated and non-denervated slice cultures were treated with Experimental procedures were performed in agreement with the recombinant TNFα (0.1 μg/ml; Sigma, Germany) for 1 or 2 days. German law on the use of laboratory animals and approved by the Soluble recombinant TNF-receptor 1 (sTNFR, 10 μg/ml, Catalog animal welfare officer of Goethe-University Frankfurt (Faculty of number: 425-R1-050, R&D systems, USA) was applied directly Medicine). Entorhino-hippocampal slice cultures were prepared after the lesion for up to 4 days (incubation medium replaced at postnatal day 4–5 as previously described (e.g., Becker et al., once with fresh sTNFR-containing medium at 2 days). 2012; Vlachos et al., 2013a). C57BL/6J and TNFα-deficient mice (and their wildtype littermates) of either sex were used (Pasparakis LASER CAPTURE MICRODISSECTION OF RE-SLICED CULTURES et al., 1996; Golan et al., 2004; obtained from Jackson Laborato- Slice cultures were washed with phosphate buffered saline (PBS; ries, USA). Slice cultures were allowed to mature for ≥18 div in 0.1 M, pH 7.4), shock frozen at −80◦C in tissue freezing medium ◦ humidified atmosphere with 5% CO2 at 35 C before experimental (Leica Microsystems, Germany), re-sliced into 10 μm thick slices assessment. on a cryostat (Leica CM 3050 S) and mounted on PET foil metal frames (Leica, Germany) as described previously (Vlachos et al., ENTORHINAL CORTEX LESION 2013b). Re-sliced cultures were fixed in ice-cold acetone for 1 min Slice cultures (18–25 div) were transected using a sterile scalpel and incubated with 0.1% toluidine blue (Merck, Germany) at blade (Figures 1A,B; e.g., Vlachos et al., 2012a, 2013a). To room temperature for 1 min, before rinsing in ultrapure water ensure complete and permanent separation of the EC from the (DNase/RNase free, Invitrogen, USA) and 70% ethanol. PET foil hippocampus, the EC was removed from the culturing dish. metal frames were mounted on a Leica DM 6000B laser capture microdissection (LMD) system (Leica Microsystems, Germany) PERFORANT PATH TRACING with the section facing downward (Burbach et al., 2003). After Anterograde tracing of the entorhino-hippocampal pathway with adjusting intensity, aperture, and cutting velocity, the pulsed ultra- biotinylated and rhodamine conjugated dextranamine Mini-Ruby violet laser beam was carefully directed along the borders of the (Figure 2A; Molecular Probes, USA) was performed as described respective hippocampal layers of interest using a 20× objective previously (Kluge et al., 1998; Prang et al., 2003; Vlachos et al., lens (Leica Laser Microdissection, Software Version 7.4.1.4853). 2012a). Tissue from the OML, the inner molecular layer (IML) and the WHOLE-CELL PATCH-CLAMP RECORDINGS granule cell layer (GCL) of the suprapyramidal blade of the DG Whole-cell voltage-clamp recordings and post hoc identification were collected. Microdissected tissue was transferred by gravity of recorded neurons were carried out as previously described into microcentrifuge tube caps placed underneath the sections, μ (Vlachos et al., 2012a). Age- and time-matched non-denervated filled with 50 l guanidine isothiocyanate (GITC)-containing cultures prepared from the same animal or littermate animals buffer (RLT Buffer, RNeasy Mini Kit, Qiagen, Germany) with served as controls. Non-denervated control (or untreated) cul- 1% ß-mercaptoethanol (AppliChem GmbH; Germany). Suc- tures were recorded alternating with the recordings of denervated cessful tissue collection was verified by visually inspecting the and/or treated cultures (c.f., Vlachos et al., 2012a). All record- content of the tube caps. All samples were frozen and stored at ◦ − ◦ ingswereperformedat35C in artificial cerebrospinal fluid 80 C. (ACSF; 126 mM NaCl, 2.5 mM KCl, 26 mM NaHCO3, 1.25 mM ISOLATING RNA AND qPCR NaH2PO4,2mMCaCl2, 2 mM MgCl2, and 10 mM glucose) ® saturated with 95% O2/5% CO2. For miniature excitatory post- RNA was isolated using the RNeasy MicroPlus Kit (Qiagen, Ger- synaptic current (mEPSC) – recordings 10 μM D-APV, 10 μM many). Purified RNA was transcribed into cDNA with the High SR-95531 and 0.5 μM TTX were added to ACSF. Patch-pipettes Capacity cDNA Reverse Transcription Kit (Applied Biosystems, contained 126 mM K-gluconate, 4 mM KCl, 4 mM ATP- USA). All kits and assays were used according to the manufac- ® Mg, 0.3 mM GTP-Na , 10 mM PO-Creatine, 10 mM HEPES turer’s instructions. The cDNA was amplified using the TaqMan 2 μ and 0.3% Biocytin (pH = 7.25 with KOH, 290 mOsm with PreAmp Master Mix Kit (Applied Biosystems, USA) using 5 l + μ sucrose). Recordings were carried out at a holding potential of PreAmp Master Mix (Applied Biosystems, USA) 2.5 l + μ −70 mV. Series resistance was monitored in 2 min intervals, and cDNA 2.5 l Assay Mix [TaqMan Gene Expression(TM)-Assay α recordings were discarded if the series resistance and leak cur- (GAPDH: 4352932E; TNF : Mm00443258_m1) from Applied ≥ ≥ Biosystems, USA] with a standard amplification protocol (14 rent changed significantly and/or reached 30 M or 50 pA, ◦ ◦ respectively. cycles: 95 C for 15 s; 60 C for 4 min). Amplified cDNAs were diluted 1:20 in ultrapure water and subjected to quantitative PCR LOCAL ELECTRICAL STIMULATION (qPCR; StepOnePlus, Applied Biosystems, USA) using a stan- A bipolar stimulation electrode (NE-200, 0.5 mm tip separation, dard amplification program (1 cycle of 50◦C for 2 min, 1 cycle Rhodes Medical Instruments, USA) was placed on the EC and of 95◦C for 10 min, 40 cycles of 95◦C for 15 s and 60◦C for current square-wave pulses (50 μA;1msat10Hz)weregen- 60 s; cut off at 36 cycles; average CT -value was: 22.7 ± 0.7 erated by a stimulus generator (STG1002 Multichannel Systems, cycles).

Frontiers in Cellular Neuroscience www.frontiersin.org December 2013 | Volume 7 | Article 257 | 2

“fncel-07-00257” — 2013/12/16 — 15:26 — page2—#2 Becker et al. TNFα and denervation-induced synaptic plasticity

FIGURE 2 | Denervation-induced homeostatic synaptic strengthening is not observed in granule cells of TNFα-deficient slice cultures at 3–4 dpl. (A) Mini-Ruby tracing of entorhino-hippocampal axons (red; ToPRO nuclear staining, blue) and electrical stimulations of the entorhinal cortex (EC) while recording evoked EPSCs from dentate FIGURE 1 | Denervation-induced homeostatic synaptic strengthening granule cells revealed an intact and functional entorhino-hippocampal α α is impaired at 3–4 dpl in the presence of soluble TNF-receptor. projection in slice cultures prepared from TNF -deficient mice (TNF -KO; (A) Schematic of an organotypic entorhino-hippocampal slice culture. three independent experiments each, up to 50 traces averaged per ± Neurons in the entorhinal cortex (EC, red) project their axons into the neuron). Evoked EPSCs (amplitude: 369 102 pA) could be blocked by μ ± outer molecular layer (OML) of the dentate gyrus (DG). By cutting the AMPA-receptor antagonist CNQX (10 M; amplitude: 9.6 2.1 pA). μ through these axons and by removing the EC from the culture dish Scale bar: 200 m. (B) Patched granule cells were filled with biocytin (plane of transection, black line), distal dendrites of dentate granule and post hoc identified using Alexa568- or Alexa488-streptavidin. Scale μ cells are denervated. IML, inner molecular layer. (B) Example of an bar: 50 m. (C–E) Whole-cell patch-clamp recordings from granule cells α entorhino-hippocampal slice culture stained with ToPRO nuclear stain of TNF -deficient slice cultures revealed an increase in the mEPSC = (blue). The white line depicts the plane of transection. Scale bar: amplitudes at 1–2 dpl but not at 3–4 dpl (n 12–16 neurons per 200 μm. (C) Sample traces of mEPSC recordings from granule cells of group, from six to eight cultures each). Data represent mean ± SEM; < non-denervated control and denervated wildtype slice cultures untreated ***p 0.001; n.s., not significant. and treated with soluble TNF-receptor (sTNFR). (D–F) While a compensatory increase in mEPSC amplitudes was observed in denervated wildtype cultures (both at 1–2 dpl and 3–4 dpl; n = 10–12 on a cryostat (Leica CM 3050 S). Sections were mounted on neurons per group, from four to five cultures each), granule cells “Superfrost plus”-microscope slides (Thermo Scientific, USA), treated with sTNFR were impaired in their ability to maintain increased mEPSC amplitudes at 3–4 dpl (n = 9–11 neurons per group, from three quenched in 0.3% (v/v)H2O2 and washed twice with PBS to four cultures each). Data represent mean ± SEM; **p < 0.01; prior to Avidin/Biotin Blocking (Invitrogen, USA). The oligo- ***p < 0.001; n.s., not significant. DNA probe (200 ng/ml) was added to the hybridization buffer [4 × SSC, 50% (v/v) Formamide, 200 mg/ml dextran sulfate sodium salt, 0.25 mg/ml ssDNA, 0.25 mg/ml tRNA, 0.01 M ◦ FLUORESCENCE IN SITU HYBRIDIZATION DTT and 1× Denhardt’s solution; at 37 C] and sections were In situ hybridization was performed using a biotin-labeled oligo- incubated with the buffer in a humidified chamber at 37◦C DNA probe (5 CT TCT CAT CCC TTT GGG GAC CGA TCA overnight. Thereafter sections were washed twice for 15 min at CC 3) directed against murine TNFα-mRNA (Fenger et al., 55◦Cin1× SSC and 0.5 × SSC [both containing 50% (v/v) 2006; Lambertsen et al., 2007). All steps were carried out under Formamide and 0.1% (v/v) Tween-20] and the formamide was RNase-free conditions. Buffers and solutions were prepared with removed by washing in Maleic acid buffer [100 mM Maleic acid, DEPC-treated water. Wildtype cultures (2 dpl) were fixed for 150 mM NaCl, 0.1% (v/v) Tween-20, pH 7.5]. The TSATM-biotin 30 min in 4% (w/v) paraformaldehyde (PFA) containing PBS system (PerkinElmer, USA) was used for signal amplification followed by 2% (w/v) PFA and 30% (w/v) sucrose in PBS at and deposited biotin was visualized by incubating the sections 4◦C overnight. Fixed cultures were snap-frozen in tissue freez- in Alexa Fluor 568 conjugated Streptavidin [1:1000 in TNB ing medium (Leica) on dry-ice and re-sliced into 18 μmsections (buffer containing 0.1 M TRIS-HCl, 0.15 M NaCl and 1%

Frontiers in Cellular Neuroscience www.frontiersin.org December 2013 | Volume 7 | Article 257 | 3

“fncel-07-00257” — 2013/12/16 — 15:26 — page3—#3 Becker et al. TNFα and denervation-induced synaptic plasticity

(v/v) provided blocking reagent) for 30 min; Molecular Probes, defined regions of interest (ROIs, 40 × 40 μm) and expressed USA]. as percent of ROI area. ROIs were positioned in the GCL (detected Counterstaining for the glial fibrillary acidic protein (GFAP) by ToPRO nuclear staining) at 0–50 μm from GCL (=IML) or was performed using a monoclonal anti-GFAP antibody (1:100 at a distance of 50–150 μm from the GCL (=OML; c.f. Vla- in TNB for 10 min; Sigma-Aldrich, Germany) followed by Alexa chos et al., 2013b). Three ROIs were analyzed per layer in each Fluor 488 conjugated mouse secondary antibody (1:50 in TNB culture and averaged values per culture were used for statisti- for 10 min, Molecular Probes, USA). Sections were washed three cal comparison. No significant difference was observed between times in PBS and mounted on microscope slides using fluorescence age- and time-matched non-denervated control cultures in these mounting medium (Dako, Denmark). experiments. Statistical comparisons were made using Mann–Whitney-test IMMUNOHISTOCHEMISTRY or Kruskal–Wallis-test followed by Dunn’s post hoc analysis. Cultures were fixed in a solution of 4% (w/v) PFA and 4% (w/v) P-values of less than 0.05 were considered a significant difference. sucroseinPBSfor1h,followedby1hin2%(w/v) PFA and 30% All values are expressed as mean ± standard error of the mean (w/v) sucrose in PBS. Fixed slice cultures were thoroughly washed, (SEM). In the figures, ∗p < 0.05, ∗∗p < 0.01, and ∗∗∗p < 0.001; not resliced into 30 μm sections (Leica VT 1000S, Germany), and significant differences are indicated with n.s. stained with antibodies against TNFα (1:500; Abcam, AB66579) and GFAP (1:2000; Sigma-Aldrich, Germany) in PBS with 10% DIGITAL ILLUSTRATIONS (v/v) normal horse serum and 0.1% (v/v) Triton X-100 based Confocal image stacks were exported as 2D-projections and stored on a previously described protocol (Vlachos et al., 2013a). Sec- as TIF files. Figures were prepared using Photoshop graphics soft- ondary antibodies (goat anti-rabbit Alexa568 and goat anti-mouse ware (Adobe, USA) and Inkscape3 (Free Software Foundation, Alexa488; Invitrogen, USA) were used at 1:2000 in PBS with USA). Image brightness and contrast were adjusted. 10% (v/v) normal horse serum and 0.1% (v/v) Triton X-100. For nuclear staining sections were incubated with To-PRO®-3 IODID RESULTS (1:5000 in PBS for 10 min; Invitrogen, USA). RECOMBINANT sTNFR IMPAIRS DENERVATION-INDUCED SYNAPTIC STRENGTHENING AT 3–4 dpl MICROSCOPY To test for the role of TNFα in denervation-induced homeo- Traced entorhino-hippocampal fibers, post hoc identified recorded static synaptic plasticity a pharmacological approach was used neurons, TNFα in situ hybridizations, and TNFα immunostain- first (Figure 1). Denervated and non-denervated wildtype cul- ings were visualized using a Nikon Eclipse C1si laser-scanning tures were treated immediately after the lesion with a recombinant microscope equipped with a 40× oil-immersion (NA 1.3, Nikon) sTNFR (10 μg/ml), which scavenges TNFα (Beattie et al., 2002; and 60× oil-immersion (NA 1.4, Nikon) objective lens. Stellwagen et al., 2005; Stellwagen and Malenka, 2006; Steinmetz and Turrigiano,2010; Santello et al.,2011). mEPSCs were recorded QUANTIFICATION AND STATISTICS from dentate granule cells at 1–2 and 3–4 dpl. During this time Electrophysiological data were analyzed using pClamp 10.2 (Axon denervation-induced homeostatic synaptic plasticity is induced Instruments, USA) and MiniAnalysis (Synaptosoft, USA) soft- and synaptic strength increases (1–2 dpl) before it reaches a plateau ware. All events were visually inspected and detected by an (3–4 dpl; Vlachos et al., 2012a). Similar to untreated denervated investigator blind to experimental condition. 250–350 events were cultures, a significant increase in mEPSC amplitudes was observed analyzed per recorded neuron. No significant differences between in sTNFR-treated wildtype cultures at 1–2 dpl (Figures 1C–F). age- and time-matched non-denervated cultures were observed However, at 3–4 dpl mEPSC amplitudes were not significantly (c.f., Vlachos et al., 2012a). Similarly no differences between 1 vs. increased in the sTNFR-treated group, while they remained high α 2 dpl, 3 vs. 4 dpl, TNF treatment for 1 vs. 2 days and sTNFR in untreated denervated cultures (mEPSC frequencies; untreated treatments for 1 vs. 2 days or 3 vs. 4 days were observed. Therefore groups, control: 1.3 ± 0.2 Hz; 1–2 dpl: 2.1 ± 0.2 Hz, n.s.; 3–4 dpl: these data were pooled (control, 1–2 dpl, 3–4 dpl, 1–2 days or 2.6 ± 0.3 Hz, n.s.; sTNFR treated groups, control: 1.2 ± 0.2 Hz; 3–4 days of treatment). 1–2 dpl: 2.4 ± 0.5 Hz, p < 0.05; 3–4 dpl: 1.2 Hz ± 0.2 Hz, Quantitative PCR data were analyzed as described by Pfaffl n.s.; mEPSC rise and decay times not significantly changed after (2001). GAPDH served as reference gene in this analysis. The entorhinal denervation in vitro). These results indicated that TNFα qPCR assay efficiency was calculated with the StepOnePlus soft- could play an important role during the late/plateau phase of ware (Applied Biosystems, USA) based on a dilution series of denervation-induced homeostatic synaptic strengthening. five samples for each assay. Data of age- and time-matched non-denervated control cultures were pooled. DENERVATION-INDUCED SYNAPTIC STRENGTHENING IS IMPAIRED AT TNFα and GFAP immunostainings were analyzed using the 3–4 dpl BUT INTACT AT 1–2 dpl IN TNFα-DEFICIENT SLICE CULTURES 1 ImageJ software package . Colocalization of TNFα and GFAP To confirm and extend these findings slice cultures prepared from 2 staining were assessed using the colocalization plugin for ImageJ . TNFα-deficient mice (Pasparakis et al., 1996; Golan et al., 2004) GFAP, TNFα and colocalized pixel-areas were determined in were used. Prior to deafferentation we verified that the DG is innervated by entorhinal fibers in these cultures by anterograde 1http://rsb.info.nih.gov/ij 2 http://rsbweb.nih.gov/ij/plugins/colocalization.html 3http://inkscape.org/

Frontiers in Cellular Neuroscience www.frontiersin.org December 2013 | Volume 7 | Article 257 | 4

“fncel-07-00257” — 2013/12/16 — 15:26 — page4—#4 Becker et al. TNFα and denervation-induced synaptic plasticity

tracing of the perforant path with Mini-Ruby (Figure 2A). Fur- thermore, by electrically stimulating the EC and simultaneously recording evoked EPSCs from dentate granule cells, we confirmed that entorhinal fibers form functional synapses with granule cells in these preparations (Figure 2A). We then performed entorhinal denervation experiments and assessed changes in excitatory synaptic strength of dentate granule cells (Figures 2C–E). Similar to our pharmacological experi- ments a significant increase in mEPSC amplitudes was observed at 1–2 dpl in TNFα-deficient preparations. However, at 3–4 dpl mEPSC amplitudes returned back to baseline in these cultures (Figures 2D,E; mEPSC frequencies; control: 3.2 ± 0.4 Hz; 1–2 dpl: 2.2 ± 0.2 Hz, n.s.; 3–4 dpl: 4.1 ± 0.5 Hz, n.s.; mEPSC rise and decay times not significantly different compared to age- and time-matched wildtype littermates). Hence, TNFα could play an important role in maintaining rather than induc- ing a homeostatic increase in excitatory synaptic strength after denervation.

TNFα-TREATMENT RESCUES THE ABILITY OF DENERVATED DENTATE GRANULE CELLS TO MAINTAIN INCREASED EXCITATORY SYNAPTIC FIGURE 3 |Tumor necrosis factor alpha (TNFα)-treatment rescues the STRENGTH AT 3–4 dpl IN TNFα-DEFICIENT SLICE CULTURES ability of denervated dentate granule cells to maintain increased α excitatory synaptic strength at 3–4 dpl inTNFα-deficient slice cultures. To test whether the impaired synaptic response in TNF - (A) Schematic illustrating the experimental design. Denervated deficient preparations at 3–4 dpl is indeed due to the lack TNFα-deficient slice cultures were treated at 2 dpl with TNFα (0.1 μg/ml) for of TNFα, another set of cultures was treated at 2 dpl 1 or 2 days and miniature excitatory postsynaptic currents (mEPSCs) were with TNFα (0.1 μg/ml) for 1 or 2 days and mEPSCs recorded (3–4 dpl). Age- and time-matched control cultures were not denervated but treated with the same concentration of TNFα for1or were recorded (Figure 3A). In these experiments a persist- 2 days. (B) Sample traces of mEPSC recordings from granule cells of ing increase in mEPSC amplitudes was detected after den- non-denervated (control) and denervated TNFα-deficient slice cultures ervation (Figures 3B–D), thus demonstrating that exoge- which were treated with TNFα. (C, D) mEPSC amplitudes were increased in α = nous TNFα-treatment restores the ability of dentate gran- denervated TNF -treated cultures (n 17 neurons; five cultures), while the same treatment had no apparent effect on mEPSC amplitudes in ule cells in TNFα-deficient preparations to maintain increased non-denervated TNFα-deficient cultures (non-denervated TNFα-treated: excitatory synaptic strength following entorhinal denervation 13.0 ± 0.7 pA; n = 8 cells; non-denervated untreated: 12.3 ± 0.8 pA; n = 12 in vitro. cells; n.s.; from three to five cultures). Results of TNFα-treated wildtype cultures are reported in the text. Data represent mean ± SEM; *p < 0.05. The same TNFα-treatment had no significant effect on mEPSC amplitudes in non-denervated. TNFα-deficient slice cultures (non-denervated untreated: 12.3 ± 0.8 pA; n = 12 cells; non-denervated TNFα-treated: 13.0 ± 0.7 pA; n = 8 test whether changes in mRNA expression reflect different cells; n.s.) and wildtype cultures (non-denervated untreated: phases after denervation, and to provide further evidence for 10.4 ± 0.3 pA; n = 10 cells; non-denervated TNFα-treated: the role of TNFα in denervation-induced synaptic plasticity 11.2 ± 0.5 pA; n = 9 cells; n.s.). Also, no significant with another technique, relative changes in TNFα-mRNA lev- increase in mEPSC amplitudes was observed when dener- els were assessed in the DG of denervated and age-matched vated wildtype cultures were treated with TNFα (denervated non-denervated wildtype cultures. Of note, our previous work untreated: 16.2 ± 0.9 pA; n = 10 cells; denervated TNFα- had revealed that a compensatory increase in excitatory synap- treated: 18.6 ± 1.0 pA; n = 17 cells; n.s.). We concluded tic strength is predominantly seen in the denervated OML from these observations that TNFα (applied at a concentra- at 3–4 dpl (Vlachos et al., 2012a). Thus, we predicted that tion of 0.1 μg/ml for 1 or 2 days to the culture medium) TNFα-mRNA levels might also be increased in the deafferented is not sufficient to induce a strengthening of excitatory post- zone. synapses in dentate granule cells. Rather, TNFα seems to To address this issue LMD was used to harvest tissue from play a role in maintaining the increased excitatory synap- the OML, the IML and the GCL at 1, 2, and 4 dpl (Figure 4A). tic strength, which is induced by entorhinal denervation The probes were obtained from the suprapyramidal blade of the in vitro. DG, i.e., the region in which we have performed our experiments. Indeed, qPCR analysis revealed an increase in TNFα-mRNA levels ENTORHINAL DENERVATION IN VITRO IS ACCOMPANIED BY CHANGES in the tissue isolated from the denervated OML at 1 dpl which IN TNFα-mRNA LEVELS IN THE DENTATE GYRUS reached the level of significance at 2 dpl and returned back to The requirement of TNFα at 3–4 dpl but not 1–2 dpl indi- baseline at 4 dpl, while in the IML no significant change was cated that denervation-induced synaptic plasticity is composed observed (Figure 4B). In the GCL TNFα-mRNA was below the of molecularly distinct (or partially overlapping) phases. To detection threshold and for this reason no quantitative analysis

Frontiers in Cellular Neuroscience www.frontiersin.org December 2013 | Volume 7 | Article 257 | 5

“fncel-07-00257” — 2013/12/16 — 15:26 — page5—#5 Becker et al. TNFα and denervation-induced synaptic plasticity

FIGURE 4 | Laser capture microdissection combined with qPCR represent mean ± SEM; **p < 0.01; n.s., not significant. (C) In reveals changes in TNFα-mRNA levels following entorhinal situ hybridization for TNFα-mRNA (red) and GFAP-immunostaining denervation in vitro. (A) Laser capture microdissection (LMD) (green) at 2 dpl. TNFα-mRNA granules colocalized with was employed to collect tissue from the inner molecular layer GFAP-immunofluorescence. While most TNFα-mRNA/GFAP-positive (IML; top right side), the outer molecular layer (OML; bottom left cells were detected in the OML (arrowheads) and some in the side), and the granule cell layer (GCL; bottom right side) of IML (arrow), only a weak TNFα-mRNA signal was present in the denervated cultures (at 1, 2, and 4 dpl) and non-denervated GCL (mostly in colocalization with GFAP). These results indicate control cultures. (B) TNFα-mRNA levels were assessed in the that astrocytes could be a major source of TNFα following isolated tissue using qPCR. An increase in TNFα-mRNA was entorhinal denervation in vitro. Scale bar: 50 μm. (D) Higher observed in the OML at 2 dpl. TNFα-mRNA levels in the IML magnification of the OML. TNFα-mRNA granules can be identified were not significantly changed. In the GCL TNFα-mRNA levels in the soma and proximal processes of GFAP-positive cells were below the detection threshold and no quantitative analysis (ToPRO nuclear staining, blue; n = 4 cultures). Scale bar: of this layer was performed (n = 4–7 cultures per group). Data 10 μm. of this layer could be performed. These results were consistent in these experiments (Figures 4C, D). Indeed, TNFα-mRNA with our earlier findings on a layer-specific homeostatic synap- was found predominantly colocalized with GFAP (Figures 4C, D). tic response of dentate granule cells (Vlachos et al., 2012a) and TNFα-mRNA granules were mostly found in somatic and peri- indicated that changes in TNFα-mRNA levels (at 1–2 dpl) may pre- somatic compartments, i.e., in proximal astrocytic processes cede the TNFα-dependent phase of denervation-induced synaptic (Figure 4D). Some TNFα-mRNA containing GFAP-positive cells plasticity. and processes were also identified in the IML and occasionally in the GCL. The majority of GFAP and TNFα-mRNA positive cells, TNFα-mRNA IS DETECTED IN ASTROCYTES FOLLOWING ENTORHINAL however, were detected in the denervated OML, consistent with DENERVATION IN VITRO our LMD-qPCR data (c.f., Figure 4B). Although these results do Previous work has demonstrated that astrocytes are activated fol- not rule out the possibility of neuronal or microglial TNFα, they lowing entorhinal denervation in the denervated zone (Rose et al., indicated that astrocytes are a source of TNFα in our experimental 1976; Steward et al., 1990, 1993; Fagan et al., 1997; Haas et al., setting. 1999; Deller et al., 2000; Liu et al., 2005; Schäfer et al., 2005). Since pharmacological blockade of network-activity leads to an ASTROCYTIC TNFα IS INCREASED IN THE DENERVATED ZONE AT 4 dpl increase in glial TNFα (Stellwagen and Malenka, 2006)wetested In light of these results we tested whether TNFα protein levels for the possibility that astrocytes could be a source of TNFα in our increase after entorhinal denervation in vitro. Cultures fixed at experiments. 2 and 4 dpl were immunostained for TNFα as well as GFAP A different set of cultures was fixed at 2 dpl and in situ and changes in immunofluorescence were determined. Consis- hybridizations for TNFα-mRNA were carried out. The slices were tent with the results obtained from in vivo entorhinal denervation counterstained for GFAP, which served as a marker for astrocytes experiments (Steward et al., 1990, 1993; Burbach et al., 2004)a

Frontiers in Cellular Neuroscience www.frontiersin.org December 2013 | Volume 7 | Article 257 | 6

“fncel-07-00257” — 2013/12/16 — 15:26 — page6—#6 Becker et al. TNFα and denervation-induced synaptic plasticity

FIGURE 5 | Astrocytes are a source of TNFα following entorhinal TNFα signal, which was colocalized with GFAP was significantly increased denervation in vitro. (A) Immunostainings for TNFα (red) and GFAP (green) in the OML at 4 dpl (C, D), suggesting that GFAP-positive astrocytes reveal that both markers are upregulated post lesion compared to control. are a major source of TNFα following entorhinal denervation in vitro Arrows are pointing at examples of TNFα and GFAP-positive cells. The (n = 6–8 cultures per group). Data represent mean ± SEM; **p < 0.01; regions of interest (ROIs) used for analysis of changes in the outer ***p < 0.001; n.s., not significant. (E) Entorhinal denervation in vitro molecular layer (OML), inner molecular layer (IML), and granule cell layer reveals that denervation-induced homeostatic synaptic plasticity can be (GCL) are indicated (ToPRO nuclear staining, blue). Scale bars: 40 μm. divided into an early synaptic strengthening response, which does not (B–D) Quantification of immunofluorescence for GFAP,TNFα colocalized require TNFα, and a late TNFα-dependent synaptic strengthening response. with GFAP and TNFα without GFAP in the GCL, IML, and OML. GFAP is TNFα plays an important role in maintaining the increased excitatory upregulated predominantly in the OML following denervation (B). Only the synaptic strength during the late-phase. significant increase in GFAP signal was observed after entorhi- reorganizationof denervated neuronal networks might be com- nal denervation in vitro (compared to non-denervated age- and posed of molecularly distinct (or partially overlapping) phases. time-matched control cultures; Figures 5A,B). We then eval- TNFα appears to play an important role during the late/plateau uated changes in TNFα immunofluorescence colocalized with phase of denervation-induced homeostatic synaptic plasticity GFAP (Figure 5C) and not colocalized with GFAP fluorescence (Figure 5E). (Figure 5D). Indeed, only the TNFα signal which colocal- ized with GFAP was significantly increased in the OML at THE ROLE OF TNFα IN DENERVATION-INDUCED HOMEOSTATIC 4 dpl. Together, these data suggested that activated astrocytes SYNAPTIC PLASTICITY are a major source of TNFα following entorhinal denervation TNFα has been implicated as one of the secreted factors, which reg- in vitro. ulate pharmacologically induced homeostatic synaptic plasticity (Stellwagen and Malenka, 2006). Evidence has been provided that DISCUSSION prolonged TTX-treatment leads to TNFα release from glia cells, The results of the present study demonstrate that TNFα is which in turn induces an increase in excitatory synaptic strength required for the maintenance (3–4 dpl) but not the induction (Stellwagen and Malenka, 2006) and also regulates inhibitory (1–2 dpl) of homeostatic synaptic plasticity following entorhinal synaptic strength (Stellwagen et al., 2005; Pribiag and Stellwagen, denervation in vitro. This finding indicates that the postlesional 2013).TheworkbySteinmetz and Turrigiano (2010) indicated

Frontiers in Cellular Neuroscience www.frontiersin.org December 2013 | Volume 7 | Article 257 | 7

“fncel-07-00257” — 2013/12/16 — 15:26 — page7—#7 Becker et al. TNFα and denervation-induced synaptic plasticity

that TNFα could have permissive rather than instructive effects in may depend on the particular experimental approach, which is TTX-induced synaptic scaling. In this earlier study (Steinmetz and used to induce homeostatic plasticity. Turrigiano, 2010) compelling experimental evidence was provided using pharmacological approaches that TNFα is necessary during DENERVATION-INDUCED HOMEOSTATIC SYNAPTIC PLASTICITY prolonged (>24 h) but not early (∼6 h) activity blockade in dis- IN VITRO IS COMPOSED OF MOLECULARLY DISTINCT PHASES sociated cortical neurons. This observation is comparable with A key finding of our study is the observation that granule cells the results of the present study, in which homeostatic synaptic of TNFα-deficient preparations are not impaired in their abil- strengthening was induced by entorhinal denervation in slice cul- ity to express homeostatic synaptic plasticity at 1–2 dpl, which tures prepared from TNFα-deficient mice. While a compensatory is supported by our pharmacological experiments in wildtype increase in mEPSC amplitudes was observed at 1–2 dpl, increased cultures in which we have used sTNFR to scavenge TNFα. This excitatory synaptic strength was not seen at 3–4 dpl in granule observation indicates that distinct signaling pathways could con- cells of TNFα-deficient cultures. Moreover, bath-application of trol different phases of denervation-induced homeostatic synaptic TNFα (0.1 μg/ml) rescued in these cultures the ability to maintain plasticity (induction vs. plateau vs. down-scaling; c.f., Vlachos increased synaptic strength at 3–4 dpl, while having no discernible et al., 2012a). Apparently, TNFα-signaling is required during effect on mEPSC amplitudes in non-denervated control cultures. the plateau-phase of denervation-induced homeostatic synaptic Thus, TNFα can be considered an important regulatory molecule, plasticity, i.e., at 3–4 dpl. Although the signals which orches- which controls the ability of denervated neurons to maintain a trate different phases of homeostatic synaptic plasticity remain homeostatic increase in excitatory synaptic strength. It will now be unknown, our LMD-qPCR data suggest that regulatory pathways important to identify the downstream signaling pathways through may act, at least in part, at the level of gene expression. Hence, a which TNFα controls this ability of neurons (for a recent review on systematic comparison of the differences in mRNA levels (and see Hulme et al., 2012) and to compare these find- protein expression at the same or later time points; Dieterich ings with data obtained in other testing conditions (c.f., review by et al., 2006; Cajigas et al., 2012) between different layers at different Santello and Volterra, 2012). points in time after entorhinal denervation (and/or pharmacolog- ical treatments) may allow to identify novel candidate regulatory ASTROCYTES ARE A SOURCE OF TNFα IN THE DENERVATED ZONE molecules involved in local and global homeostatic synaptic In our earlier work we reported evidence that excitatory plasticity. synaptic strength is predominantly increased in the dener- vated layer at 3–4 dpl (Vlachos et al., 2012a). This finding is THE ROLE OF TNFα-DEPENDENT HOMEOSTATIC SYNAPTIC PLASTICITY in line with the layer-specific increase in TNFα-mRNA lev- IN NEUROLOGICAL DISEASES els in the OML at 2 dpl, as revealed by a combination of Since entorhinal denervation was employed in our study to induce LMD and qPCR. Using in situ hybridization the TNFα-mRNA homeostatic synaptic strengthening it is conceivable that TNFα- signal could be localized to GFAP-positive cells, i.e., acti- mediated homeostatic synaptic plasticity could be of relevance vated astrocytes. Immunostaining for TNFα and GFAP fur- for a broad range of neurological diseases, which are accompa- ther corroborated these findings and revealed an enrichment nied by a deafferentation of neurons. In fact, several neurological of TNFα in GFAP-positive astrocytes in the denervated zone diseases are associated with increased TNFα-expression levels at 4 dpl. Taken together, we conclude that astrocytes are a (Sriram and O’Callaghan, 2007). While it has been proposed major source of TNFα in our experimental setting, although that the pathological release of TNFα could lead to alterations this finding does not rule out a contribution from other TNFα in physiological network functions (Small, 2008; Park et al., 2010; sources, e.g., microglia (Lambertsen et al., 2001; Fenger et al., Montgomery and Bowers, 2012), the biological consequences of 2006)orneurons(Harry et al., 2008; Janelsins et al., 2008) TNFα-mediated homeostatic synaptic plasticity remain unclear. in vivo. On the one hand the maintenance of homeostatic synaptic Astrocytes activated by entorhinal denervation delineate the plasticity could be beneficial by promoting stability during the denervated OML of the DG (Fagan et al., 1997; Haas et al., 1999; post-lesional reorganization of denervated networks. On the other Deller et al., 2000; Liu et al., 2005; Schäfer et al., 2005). They are hand a persisting increase in excitatory synaptic strength could also thus in an ideal spatial position to regulate the remodeling occur- aggravate the susceptibility for runaway excitation and epileptic ring within the denervated layer. Based on the findings of this discharges in lesioned neuronal networks (Savin et al., 2009). It study, we propose that activated astrocytes could play an impor- therefore remains to be shown whether targeting TNFα-signaling tant role in the regulation of functional changes after entorhinal could have beneficial or detrimental effects for the course of a neu- denervation. Their presence in the OML could lead to a longer rological disease (Tobinick et al., 2006; McCoy and Tansey, 2008; lasting homeostatic increase in the strength of surviving excita- Zhao et al., 2011). tory synapses, which could uphold the level of excitatory drive to a denervated granule cell, e.g., until this cell is reinnervated THE ROLE OF TNFα IN DENERVATION-INDUCED STRUCTURAL by sprouting excitatory axons. Whether astrocyte-derived TNFα REMODELING is also required in other testing conditions in which local home- In one of our recent studies (Vlachos et al., 2013a)wewereableto ostatic responses have been reported (Sutton et al., 2006; Branco provide experimental evidence which suggests that denervation- et al., 2008; Hou et al., 2008; Kim and Tsien,2008; Deeg and Aizen- induced homeostatic synaptic plasticity can lead to heterosynaptic man, 2011; Mitra et al., 2012) remains unknown at present and competition between strengthened excitatory synapses and newly

Frontiers in Cellular Neuroscience www.frontiersin.org December 2013 | Volume 7 | Article 257 | 8

“fncel-07-00257” — 2013/12/16 — 15:26 — page8—#8 Becker et al. TNFα and denervation-induced synaptic plasticity

formed dendritic spines, i.e., spines formed after denervation. Fenger, C., Drojdahl, N., Wirenfeldt, M., Sylvest, L., Jorgensen, O. S., Meldgaard, M., This mechanism seems to delay spine density recovery after den- et al. (2006). Tumor necrosis factor and its p55 and p75 receptors are not required ervation via the destabilization of new spines (Vlachos et al., for axonal lesion-induced microgliosis in mouse fascia dentata. Glia 54, 591–605. doi: 10.1002/glia.20405 2013a). Noteworthy, spine destabilization was observed in the Golan, H., Levav, T., Mendelsohn, A., and Huleihel, M. (2004). Involvement of denervated zone (Vlachos et al., 2012b, 2013a) as was also the tumor necrosis factor alpha in hippocampal development and function. Cereb. case for homeostatic synaptic strengthening at 3–4 dpl (Vlachos Cortex 14, 97–105. doi: 10.1093/cercor/bhg108 et al., 2012a) and the increase of TNFα-mRNA levels and TNFα- Haas, C. A., Rauch, U., Thon, N., Merten, T., and Deller, T. (1999). Entorhinal immunofluorescence signal (this study). Hence, it will now be cortex lesion in adult rats induces the expression of the neuronal chondroitin α sulfate proteoglycan neurocan in reactive astrocytes. J. Neurosci. 19, 9953– interesting to study the effects of TNF on spine numbers and 9963. dynamics following entorhinal denervation in vitro and to assess Harry, G. J., Lefebvre D’Hellencourt, C., McPherson, C. A., Funk, J. A., Aoyama, M., whether inhibition of TNFα-signaling accelerates spine density and Wine, R. N. (2008). Tumor necrosis factor p55 and p75 receptors are involved recovery after denervation. in chemical-induced apoptosis of dentate granule neurons. J. Neurochem. 106, 281–298. doi: 10.1111/j.1471-4159.2008.05382.x ACKNOWLEDGMENTS He, P., Liu, Q., Wu, J., and Shen, Y. (2012). Genetic deletion of TNF receptor suppresses excitatory synaptic transmission via reducing AMPA receptor synap- We thank C. Nolte-Uhl for her skillful assistance in tissue culturing tic localization in cortical neurons. FASEB J. 26, 334–345. doi: 10.1096/fj.11- and LMD-qPCR experiments. Wealso thank Dr. J. Schlaudraff and 192716 L. M. Willems for their help in data analysis and Drs. N. Maggio Hou, Q., Zhang, D., Jarzylo, L., Huganir, R. L., and Man, H. Y. (2008). and D. Del Turco for helpful discussions. Supported by Deutsche Homeostatic regulation of AMPA receptor expression at single hippocampal Forschungsgemeinschaft (CRC 1080). synapses. Proc. Natl. Acad. Sci. U.S.A. 105, 775–780. doi: 10.1073/pnas.0706 447105 Hulme, S. R., Jones, O. D., Ireland, D. R., and Abraham, W. C. (2012). Calcium- REFERENCES dependent but action potential-independent BCM-like metaplasticity in the Beattie, E. C., Stellwagen, D., Morishita, W., Bresnahan, J. C., Ha, B. K., Von Zastrow, hippocampus. J. Neurosci. 32, 6785–6794. doi: 10.1523/JNEUROSCI.0634- M., et al. (2002). Control of synaptic strength by glial TNFalpha. Science 295, 12.2012 2282–2285. doi: 10.1126/science.1067859 Janelsins, M. C., Mastrangelo, M. A., Park, K. M., Sudol, K. L., Narrow, W. C., Becker, D., Willems, L. M., Vnencak, M., Zahn, N., Schuldt, G., Jedlicka, P., Oddo, S., et al. (2008). Chronic neuron-specific tumor necrosis factor-alpha et al. (2012). Functional and structural properties of dentate granule cells with expression enhances the local inflammatory environment ultimately leading hilar basal dendrites in mouse entorhino-hippocampal slice cultures. PLoS ONE to neuronal death in 3xTg-AD mice. Am. J. Pathol. 173, 1768–1782. doi: 7:e48500. doi: 10.1371/journal.pone.0048500 10.2353/ajpath.2008.080528 Branco, T., Staras, K., Darcy, K. J., and Goda, Y. (2008). Local dendritic activ- Kim, J., and Tsien, R. W. (2008). -specific adaptations to inactivity in hip- ity sets release probability at hippocampal synapses. Neuron 59, 475–485. doi: pocampal circuits achieve homeostatic gain control while dampening network 10.1016/j.neuron.2008.07.006 reverberation. Neuron 58, 925–937. doi: 10.1016/j.neuron.2008.05.009 Burbach, G. J., Dehn, D., Del Turco, D., and Deller, T. (2003). Quantification of layer- Kluge, A., Hailer, N. P., Horvath, T. L., Bechmann, I., and Nitsch, R. (1998). Tracing specific gene expression in the hippocampus: effective use of laser microdissection of the entorhinal-hippocampal pathway in vitro. Hippocampus 8, 57–68. doi: in combination with quantitative RT-PCR. J. Neurosci. Methods 131, 83–91. doi: 10.1002/(SICI)1098-1063(1998)8:1<57::AID-HIPO6>3.0.CO;2-4 10.1016/S0165-0270(03)00232-2 Lambertsen, K. L., Clausen, B. H., Fenger, C., Wulf, H., Owens, T., Dagnaes-Hansen, Burbach, G. J., Dehn, D., Del Turco, D., Staufenbiel, M., and Deller, T. (2004). F., et al. (2007). Microglia and macrophages express tumor necrosis factor recep- Laser microdissection reveals regional and cellular differences in GFAP mRNA tor p75 following middle cerebral artery occlusion in mice. Neuroscience 144, upregulation following brain injury, axonal denervation, and amyloid plaque 934–949. doi: 10.1016/j.neuroscience.2006.10.046 deposition. Glia 48, 76–84. doi: 10.1002/glia.20057 Lambertsen, K. L., Gregersen, R., Drojdahl, N., Owens, T., and Finsen, B. (2001). Cajigas, I. J., Tushev, G., Will, T. J., Tom Dieck, S., Fuerst, N., and Schu- A specific and sensitive method for visualization of tumor necrosis factor in the man, E. M. (2012). The local transcriptome in the synaptic neuropil revealed murine central nervous system. Brain Res. Brain Res. Protoc. 7, 175–191. doi: by deep sequencing and high-resolution imaging. Neuron 74, 453–466. doi: 10.1016/S1385-299X(01)00062-9 10.1016/j.neuron.2012.02.036 Leonoudakis, D., Zhao, P., and Beattie, E. C. (2008). Rapid tumor necrosis factor Davis, G. W. (2013). Homeostatic signaling and the stabilization of neural function. alpha-induced exocytosis of glutamate receptor 2-lacking AMPA receptors to Neuron 18, 3. extrasynaptic plasma membrane potentiates excitotoxicity. J. Neurosci. 28, 2119– Deeg, K. E., and Aizenman, C. D. (2011). Sensory modality-specific homeostatic 2130. doi: 10.1523/JNEUROSCI.5159-07.2008 plasticity in the developing optic tectum. Nat. Neurosci. 14, 548–550. doi: Liu, X., Ying, G., Wang, W., Dong, J., Wang, Y., Ni, Z., et al. (2005). Entorhinal 10.1038/nn.2772 deafferentation induces upregulation of SPARC in the mouse hippocampus. Brain Deller, T., Haas, C. A., and Frotscher, M. (2000). Reorganization of the rat fas- Res. Mol. Brain Res. 141, 58–65. doi: 10.1016/j.molbrainres.2005.08.003 cia dentata after a unilateral entorhinal cortex lesion. Role of the extracellular McCoy, M. K., and Tansey, M. G. (2008). TNF signaling inhibition in the matrix. Ann. N. Y. Acad. Sci. 911, 207–220. doi: 10.1111/j.1749-6632.2000.tb CNS: implications for normal brain function and neurodegenerative disease. J. 06728.x Neuroinflammation 5, 45. doi: 10.1186/1742-2094-5-45 Del Turco, D., and Deller, T. (2007). Organotypic entorhino-hippocampal slice Mitra, A., Mitra, S. S., and Tsien, R. W. (2012). Heterogeneous reallocation of cultures–atooltostudythemolecular and cellular regulation of axonal regen- presynaptic efficacy in recurrent excitatory circuits adapting to inactivity. Nat. eration and collateral sprouting in vitro. Methods Mol. Biol. 399, 55–66. doi: Neurosci. 15, 250–257. doi: 10.1038/nn.3004 10.1007/978-1-59745-504-6_5 Montgomery, S. L., and Bowers, W. J. (2012). Tumor necrosis factor-alpha and the Dieterich, D. C., Link, A. J., Graumann, J., Tirrell, D. A., and Schuman, E. roles it plays in homeostatic and degenerative processes within the central nervous M. (2006). Selective identification of newly synthesized proteins in mam- system. J. Neuroimmune Pharmacol. 7, 42–59. doi: 10.1007/s11481-011-9287-2 malian cells using bioorthogonal noncanonical amino acid tagging (BON- Müller, C. M., Vlachos, A., and Deller, T. (2010). Calcium of acutely CAT). Proc. Natl. Acad. Sci. U.S.A. 103, 9482–9487. doi: 10.1073/pnas.06016 denervated and lesioned dentate gyrus in organotypic entorhino-hippocampal 37103 co-cultures. Cell Calcium 47, 242–252. doi: 10.1016/j.ceca.2009.12.006 Fagan, A. M., Suhr, S. T., Lucidi-Phillipi, C. A., Peterson, D. A., Holtzman, D. M., and Park, K. M., Yule, D. I., and Bowers, W. J. (2010). Impaired TNF-alpha control of Gage, F. H. (1997). Endogenous FGF-2 is important for cholinergic sprouting in IP3R-mediated Ca2+ release in Alzheimer’s disease mouse neurons. Cell Signal the denervated hippocampus. J. Neurosci. 17, 2499–2511. 22, 519–526. doi: 10.1016/j.cellsig.2009.11.006

Frontiers in Cellular Neuroscience www.frontiersin.org December 2013 | Volume 7 | Article 257 | 9

“fncel-07-00257” — 2013/12/16 — 15:26 — page9—#9 Becker et al. TNFα and denervation-induced synaptic plasticity

Pasparakis, M., Alexopoulou, L., Episkopou, V., and Kollias, G. (1996). Immune Sutton, M. A., Ito, H. T., Cressy, P., Kempf, C., Woo, J. C., and Schuman, and inflammatory responses in TNF alpha-deficient mice: a critical require- E. M. (2006). Miniature neurotransmission stabilizes synaptic function via ment for TNF alpha in the formation of primary B cell follicles, follicular tonic suppression of local dendritic protein synthesis. Cell 125, 785–799. doi: dendritic cell networks and germinal centers, and in the maturation of the 10.1016/j.cell.2006.03.040 humoral immune response. J. Exp. Med. 184, 1397–1411. doi: 10.1084/jem.184. Tobinick, E., Gross, H., Weinberger, A., and Cohen, H. (2006). TNF-alpha 4.1397 modulation for treatment of Alzheimer’s disease: a 6-month pilot study. Pfaffl, M. W. (2001). A new mathematical model for relative quantification in real- MedGenMed. 8, 25. time RT-PCR. Nucleic Acids Res. 29, e45. doi: 10.1093/nar/29.9.e45 Turrigiano, G. (2012). Homeostatic synaptic plasticity: local and global mechanisms Prang, P., Del Turco, D., and Deller, T. (2003). Associational sprouting in the for stabilizing neuronal function. Cold Spring Harb. Perspect. Biol. 4, a005736. doi: mouse fascia dentata after entorhinal lesion in vitro. Brain Res. 978, 205–212. 10.1101/cshperspect.a005736 doi: 10.1016/S0006-8993(03)02836-1 Vitureira, N., Letellier, M., and Goda, Y. (2012). Homeostatic synaptic plasticity: Pribiag, H., and Stellwagen, D. (2013). TNF-alpha downregulates inhibitory from single synapses to neural circuits. Curr. Opin. Neurobiol. 22, 516–521. doi: neurotransmission through protein phosphatase 1-dependent trafficking of 10.1016/j.conb.2011.09.006 GABAA receptors. J. Neurosci. 33, 15879–15893. doi: 10.1523/JNEUROSCI.0530- Vlachos, A., Becker, D., Jedlicka, P., Winkels, R., Roeper, J., and Deller, T. (2012a). 13.2013 Entorhinal denervation induces homeostatic synaptic scaling of excitatory post- Rose, G., Lynch, G., and Cotman, C. W. (1976). Hypertrophy and redistribution synapses of dentate granule cells in mouse organotypic slice cultures. PLoS ONE of astrocytes in the deafferented dentate gyrus. Brain Res. Bull. 1, 87–92. doi: 7:e32883. doi: 10.1371/journal.pone.0032883 10.1016/0361-9230(76)90052-6 Vlachos, A., Orth, C. B., Schneider, G., and Deller, T. (2012b). Time-lapse imaging Santello, M., Bezzi, P., and Volterra, A. (2011). TNFalpha controls glutamatergic of granule cells in mouse entorhino-hippocampal slice cultures reveals changes gliotransmission in the hippocampal dentate gyrus. Neuron 69, 988–1001. doi: in spine stability after entorhinal denervation. J. Comp. Neurol. 520, 1891–1902. 10.1016/j.neuron.2011.02.003 doi: 10.1002/cne.23017 Santello, M., and Volterra, A. (2012). TNFalpha in synaptic function: switching Vlachos, A., Helias, M., Becker, D., Diesmann, M., and Deller, T. (2013a). NMDA- gears. Trends Neurosci. 35, 638–647. doi: 10.1016/j.tins.2012.06.001 receptor inhibition increases spine stability of denervated mouse dentate granule Savin, C., Triesch, J., and Meyer-Hermann, M. (2009). Epileptogenesis due cells and accelerates spine density recovery following entorhinal denervation in to glia-mediated synaptic scaling. J. R. Soc. Interface 6, 655–668. doi: vitro. Neurobiol. Dis. 59, 267–276. doi: 10.1016/j.nbd.2013.07.018 10.1098/rsif.2008.0387 Vlachos, A., Ikenberg, B., Lenz, M., Becker, D., Reifenberg, K., Bas-Orth, Schäfer, M., Brauer, A. U., Savaskan, N. E., Rathjen, F. G., and Brummendorf, C., et al. (2013b). Synaptopodin regulates denervation-induced homeostatic T. (2005). Neurotractin/kilon promotes neurite outgrowth and is expressed on synaptic plasticity. Proc. Natl. Acad. Sci. U.S.A. 110, 8242–8247. doi: reactive astrocytes after entorhinal cortex lesion. Mol. Cell. Neurosci. 29, 580–590. 10.1073/pnas.1213677110 doi: 10.1016/j.mcn.2005.04.010 Wang, G., Gilbert, J., and Man, H. Y. (2012). AMPA receptor trafficking in home- Small, D. H. (2008). Network dysfunction in Alzheimer’s disease: does synap- ostatic synaptic plasticity: functional molecules and signaling cascades. Neural tic scaling drive disease progression? Trends Mol. Med. 14, 103–108. doi: Plast. 2012, 825364. doi: 10.1155/2012/825364 10.1016/j.molmed.2007.12.006 Zhao, J., O’Connor, T., and Vassar, R. (2011). The contribution of activated astro- Sriram, K., and O’Callaghan, J. P. (2007). Divergent roles for tumor necrosis factor- cytes to Abeta production: implications for Alzheimer’s disease pathogenesis. J. alpha in the brain. J. Neuroimmune Pharmacol. 2, 140–153. doi: 10.1007/s11481- Neuroinflammation 8, 150. doi: 10.1186/1742-2094-8-150 007-9070-6 Steinmetz, C. C., and Turrigiano, G. G. (2010). Tumor necrosis factor- Conflict of Interest Statement: The authors declare that the research was conducted alpha signaling maintains the ability of cortical synapses to express synap- in the absence of any commercial or financial relationships that could be construed tic scaling. J. Neurosci. 30, 14685–14690. doi: 10.1523/JNEUROSCI.2210- as a potential conflict of interest. 10.2010 Stellwagen, D., Beattie, E. C., Seo, J. Y., and Malenka, R. C. (2005). Differential Received: 28 August 2013; accepted: 26 November 2013; published online: 18 December regulation of AMPA receptor and GABA receptor trafficking by tumor necrosis 2013. factor-alpha. J. Neurosci. 25, 3219–3228. doi: 10.1523/JNEUROSCI.4486-04.2005 Citation: Becker D, Zahn N, Deller T and Vlachos A (2013) Tumor necro- Stellwagen, D., and Malenka, R. C. (2006). Synaptic scaling mediated by glial TNF- sis factor alpha maintains denervation-induced homeostatic synaptic plasticity of alpha. Nature 440, 1054–1059. doi: 10.1038/nature04671 mouse dentate granule cells. Front. Cell. Neurosci. 7:257. doi: 10.3389/fncel.2013. Steward, O., Kelley, M. S., and Torre, E. R. (1993). The process of reinnervation in the 00257 dentate gyrus of adult rats: temporal relationship between changes in the levels This article was submitted to the journal Frontiers in Cellular Neuroscience. of glial fibrillary acidic protein (GFAP) and GFAP mRNA in reactive astrocytes. Copyright © 2013 Becker, Zahn, Deller and Vlachos. This is an open-access article Exp. Neurol. 124, 167–183. doi: 10.1006/exnr.1993.1187 distributed under the terms of the Creative Commons Attribution License (CC BY). Steward, O., Torre, E. R., Phillips, L. L., and Trimmer, P. A. (1990). The The use, distribution or reproduction in other forums is permitted, provided the original process of reinnervation in the dentate gyrus of adult rats: time course of author(s) or licensor are credited and that the original publication in this journal is cited, increases in mRNA for glial fibrillary acidic protein. J. Neurosci. 10, 2373– in accordance with accepted academic practice. No use, distribution or reproduction is 2384. permitted which does not comply with these terms.

Frontiers in Cellular Neuroscience www.frontiersin.org December 2013 | Volume 7 | Article 257 | 10

“fncel-07-00257” — 2013/12/16 — 15:26 — page 10 — #10